ROIs in ENVI are stored in pixel-based co-ordinates, with no geographic information at all. The size of the image used to create the ROI is stored in the ROI file, and when you open the ROI list for an image it only displays the currently loaded ROIs that match the image dimensions.
Although your images are all of the same location, in the same projection and with the same datum, they may not be exactly the same size. The Landsat 8 image, for example, will almost certain have a different size - as the resolution for some Landsat 8 bands has changed.
Luckily, there is a simple way to solve this: however, it relies on your images being accurately georeferenced. Trying loading the images into separate ENVI displays, and then create a geographic link between the windows (Tools->Link->Geographic Link). You should then be able to move around within the images and check that they line up properly. If they don't then you'll need to do some image registration (see the Map->Registration menu).
Once you've made sure they're correctly lined up, you can convert your ROIs from one image to be used in the others. To do this, open your Landsat 7 or Landsat 8 image, go to Tools->ROIs->Reconcile ROIs via map and choose the ROI, and it should convert the ROI for use within the Landsat 7 image. Again, by doing a geographic link between the images you can check that the ROI is in the right place.
Best Answer
There are multiple "gap fill" techniques which essentially take multiple L7-SLC off scenes and combines them to create a gap free image. This may or may not work for you application. For land classification it should be acceptable, assuming you can find other L7 scenes of a similar vintage.
You should be very aware of what these procedures are doing and be sure they are acceptable for your use case. Generally, this is better than any interpolation method since the "gap fill" pixels would simply be older, not entirely made up.
If you are working with crop classification this may not be acceptable, since even a few days can significantly alter the signature of vegetation.
The USGS has a great starting point for "gap fill" procedures http://landsat.usgs.gov/using_Landsat_7_data.php.
Tools such as ERDAS Imagine and ENVI have native tools to tackle this problem.
-Tim